KernelScan.io

HIGH

accel/amdxdna HwCtx UAF

CVE-2026-45980

CVSS 7.8 / 10.0 NVD

CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H

KernelScan AI7.8HIGH

01

In the Linux kernel, the following vulnerability has been resolved: accel/amdxdna: Stop job scheduling across aie2_release_resource() Running jobs on a hardware context while it is in the process of releasing resources can lead to use-after-free and crashes. Fix this by stopping job scheduling before calling aie2_release_resource() and restarting it after the release completes. Additionally, aie2_sched_job_run() now checks whether the hardware context is still active.

02

Engine v0.2.0

Risk summary

Local users with access to AMD XDNA AI accelerator hardware can trigger use-after-free conditions during hardware context destruction, potentially leading to kernel crashes or privilege escalation. The vulnerability affects systems with AMD AI accelerator cards and requires local access with device permissions.

Affecteddrivers/accel/amdxdna/aie2_ctx.c (AMD XDNA accelerator)

Vulnerability analysis

The root cause is a race condition where job scheduling continues to run on a hardware context while aie2_release_resource() is destroying that context's resources. Jobs can access freed memory structures, causing use-after-free conditions. The fix implements proper synchronization by stopping the DRM scheduler before resource release, adding a status check in aie2_sched_job_run() to prevent execution on inactive contexts, and restarting the scheduler after cleanup completes. This requires local access to the AMD XDNA accelerator device file.

03

BranchFixed inPatch commit
6.156.15b79d31dce49b
6.186.18.14688c3ff079b1
6.196.19.4f1370241fe80
mainline7.0